What problem are we trying to solve here? At this point in time, guided navigation with completely offline maps and GPS has already been a no-brainer off-the-shelf thing for decades.
AFAIK it's almost always enhanced by things cell tower data, wifi network data, and external data sources (besides the satellites). Look up GPS/GNSS enhancement and augmentation for the latter.
Not GPS and WAAS. Not GPS and RTK. Not GPS + wifi + BT + cellular. I didn't mean any of those things, so I did not write any of those things.
If the thing is more than GPS -- by itself -- then that's outside the scope of what I was referring to with the juxtaposition of the words "GPS" and "itself".
(If a thing -- by itself -- can be better specified to be that way using concise phrasing, then I'm all ears.)
You could get more accurate fix with RTK data, but I'm not sure if that's actually widely used. And in any case that doesn't require active communications either, you could get correction data from satellite broadcasts too.
Technically it only requires an antenna that can listen on the LTE band (or even GSM). Trilaterating based on cell towers with a hackRF or other SDR is a fun exercise.
If your device has zero GPS signal then you can get ~100m accuracy from the cellular signals alone. If your device doesn't have "enhanced GPS" then you can get ~1m accuracy from the GPS signals alone.
Note that this changed with 5G beamforming. The new towers have a much better idea of where you are. (My understanding (thanks to other HN commenters) is that technically it's possible to do beamforming without deriving precise 3D coordinates but that this isn't how it's done in practice.)